The objective of this study was to further clarify the cancer risk associated with asbestos exposure in railway carriage construction and repair. The cohort included 734 subjects employed between 1 January 1945 and 31 December 1969. Vital status was ascertained at 31 December 1997. Mortality was investigated in the time span 1970-97. Forty-two subjects (6%) were lost to follow-up and eight causes of death (4%) could not be ascertained. The overall mortality was not above the expected value. Among neoplastic diseases, excesses were observed for lung ştandardized mortality ratio (SMR) = 124; 90% confidence interval (CI) = 87-172; 26 obs), pleura (SMR = 1,327; CI = 523-2,790; 5 obs), larynx (SMR = 240; CI = 95-505; 5 obs), liver (SMR = 241; CI = 126-420; 9 obs), pancreas (SMR = 224; CI = 98-443; 6 obs) and multiple myeloma (SMR = 429; CI = 117-1,109; 3 obs). The observed excess of lung and pleural neoplasms can be causally related to asbestos exposure in the manufacture of railway carriages. A causal role of asbestos exposure in the raised SMRs from laryngeal and pancreatic neoplasms and multiple myeloma cannot be conclusively proven.